Complete plastid genome sequences of two species of the Neotropical genus Brunellia (Brunelliaceae) Janice Valencia-D1, José Murillo-A2, Clara Inés Orozco2, Carlos Parra-O2 and Kurt M. Neubig1 1 School of Biological Sciences, Southern Illinois University at Carbondale, Carbondale, IL, United States of America 2 Instituto de Ciencias Naturales, Universidad Nacional de Colombia, Bogotá D.C., Colombia ABSTRACT Here we present the first two complete plastid genomes for Brunelliaceae, a Neotropical family with a single genus, Brunellia. We surveyed the entire plastid genome in order to find variable cpDNA regions for further phylogenetic analyses across the family. We sampled morphologically different species, B. antioquensis and B. trianae, and found that the plastid genomes are 157,685 and 157,775 bp in length and display the typical quadripartite structure found in angiosperms. Despite the clear morphological distinction between both species, the molecular data show a very low level of divergence. The amount of nucleotide substitutions per site is one of the lowest reported to date among published congeneric studies (π D 0:00025). The plastid genomes have gene order and content coincident with other COM (Celastrales, Oxalidales, Malpighiales) relatives. Phylogenetic analyses of selected superrosid representatives show high bootstrap support for the ((C,M)O) topology. The N-fixing clade appears as the sister group of the COM clade and Zygophyllales as the sister to the rest of the fabids group. Subjects Bioinformatics, Genomics, Plant Science, Forestry Keywords Brunellia, Brunelliaceae, COM group, Fabids, Oxalidales, Plastid genome, Plastome, Submitted 23 July 2019 Rosid clade, Andean trees Accepted 13 December 2019 Published 28 January 2020 Corresponding author INTRODUCTION Janice Valencia-D, Brunellia Ruiz & Pav., with about 60 species, is the only genus within the Neotropical family [email protected] Brunelliaceae Engl. Its range of distribution is from southern Mexico to Bolivia, and a single Academic editor Tatiana Tatarinova species reaches the Greater Antilles (Orozco et al., 2017). They are evergreen trees, some of which can reach 40 m in height and 1 m DBH in high mountain forests, or grow less Additional Information and Declarations can be found on than 10 m high at over 3,600 m of elevation (Cuatrecasas, 1970). They are also important page 12 elements of the Andean ecosystems where they have diversified, and comprise part of the DOI 10.7717/peerj.8392 high endemism of these zones (Orozco, 2001). The narrow range of distribution of the vast Copyright majority of the species make Brunellia an interesting group to study from phylogenetic and 2020 Valencia-D et al. biogeographical perspectives. Distributed under Brunelliaceae belong to the order Oxalidales in a clade formed with Cephalotaceae, Creative Commons CC-BY 4.0 Cunoniaceae and Elaeocarpaceae (Zhang & Simmons, 2006; Soltis et al., 2011; Heibl & OPEN ACCESS Renner, 2012; Sun et al., 2016a; Li et al., 2019). Oxalidales belong to the commonly called How to cite this article Valencia-D J, Murillo-A J, Parra-O C, Neubig KM. 2020. Complete plastid genome sequences of two species of the Neotropical genus Brunellia (Brunelliaceae). PeerJ 8:e8392 http://doi.org/10.7717/peerj.8392 `COM clade' together with Celastrales and Malpighiales (Matthews & Endress, 2006), and it has been broadly recognized as monophyletic based on plastid data (but see Zhao et al., 2016; Zeng et al., 2017; Leebens-Mack et al., 2019). However, despite support for monophyly of each of these three orders, there is still some controversy regarding the relationships among them. For Brunelliaceae, the first phylogenetic analysis was made by Orozco (2001) based on morphological characters. In that study, Brunellia was recognized as monophyletic with five morphological characters as synapomorphies. Orozco (2001) proposed five sections for the genus, but the phylogenetic tree had poor resolution and low support. A phylogenetic analysis of Brunellia using molecular characters, both from nuclear and plastid DNA is currently ongoing (José Murillo-A, Clara Inés Orozco, Carlos Parra-O, Alvaro J. Perez and Katya Romoleroux, 2019, unpublished data). In a preliminary survey, plastid DNA regions (atpB-matK, ndhF, psaB-rps14, psaI-accD, psbJ-petA, psbA-trnH, rbcL, trnS-trnG) were used, but they showed low sequence variation, making them unsuitable for resolving relationships within this genus. With the purpose of identifying variable cpDNA regions, we surveyed the entire plastid genome of two morphologically different species, Brunellia antioquensis (Cuatrec.) Cuatrec. and B. trianae Cuatrec., which belong to the biggest sections in the genus, Sect. Brunellia and Sect. Simplicifolia, respectively. Brunellia antioquensis has an ochraceous to fulvous pubescence, stipules 3–5 mm long, compound leaves and fruits with U-shaped endocarp, whereas B. trianae has an appressed, lanate to arachnoid, ochraceous indument, stipules 0.5–2 mm long, simple leaves and fruits with boat-shaped endocarp (Fig. 1). Here we explore the variability and utility of the plastid DNA in Brunelliaceae and document for first-time complete cp-genome sequences for two species. We characterize the plastid genome of each species and compare them in terms of divergence hotspots in coding and non-coding regions. We also analyze the plastid gene organization of Brunellia and ten other COM representatives. Finally, using 75 protein-coding regions available for 41 selected superrosid species; we establish the phylogenetic position of Brunelliaceae and produce a hypothesis of relationships among the COM clade and related orders. MATERIALS & METHODS DNA sampling, extraction and sequencing The samples used in this study were collected under the institutional Universidad Nacional de Colombia collection permit (Number: 0255 March 2014). Collections were made in Colombia in Cerro del Padre Amaya, Antioquia department, rural areas of Medellín, in October 2012 (B. antioquensis C.I. Orozco 4001, B. trianae C.I. Orozco 4015). Sections of leaves were dried on silica gel. Vouchers specimens of both collections were deposited at the Herbario Nacional Colombiano (COL). These species belong to two different clades that were identified using ITS and ETS regions (José Murillo-A, Clara Inés Orozco, Carlos Parra-O, Alvaro J. Perez and Katya Romoleroux, pers. obs., 2019). DNA was extracted using a CTAB method (Doyle & Doyle, 1987), followed by a silica purification column step and elution in Tris-EDTA (Neubig et al., 2014). DNA samples Valencia-D. et al. (2020), PeerJ, DOI 10.7717/peerj.8392 2/19 Figure 1 Brunellia antioquensis. (Cuatrec.) Cuatrec. and B. trianae Cuatrec. Brunellia antioquensis: (A) habit, (B) twig with infructescences; (C) young leaflets and stipels. Brunellia trianae: (D) twig with adult leaves and vegetative buds; (E) twig with infructescences. (A)–(C) from Orozco et al. 4021 (COL), (D)–(E) from Orozco et al. 4019 (COL). Scale for (A), 2 m; (B), 2 cm; (C), 1 cm; (D)–(E), 2 cm. Photos by Clara I. Orozco. Full-size DOI: 10.7717/peerj.8392/fig-1 were adjusted to 50 ng/mL to be sheared to fragments of approximately 500 bp. Library preparation, barcoding and sequencing on an Illumina HiSeqX were conducted at Rapid Genomics LLC (Gainesville, FL, USA). Plastid genome assembly and annotation Sequencing process produced, on average, 7.5 Gb (s.d. D 0.2 Gb) reads per sample. Paired- end reads were trimmed (quality at 0.05 probability) and later assembled in Geneious 10.2.3 Valencia-D. et al. (2020), PeerJ, DOI 10.7717/peerj.8392 3/19 (Biomatters Ltd, Auckland, New Zealand). The assembly was performed by a combination of reference and de novo assemblies using as a reference the plastid genome of Averrhoa carambola, Oxalidaceae (GenBank accession KX364202). Annotations of gene-regions and rRNAs were transferred in Geneious from Averrhoa reference. The tRNAs were annotated using tRNAscan-SE v2.0 (Lowe & Chan, 2016) implemented in CHLOROBOX (https://chlorobox.mpimp-golm.mpg.de/geseq.html). All annotations were manually reviewed and, if necessary, edited. The circular map of the plastid genome was generated with OGDRAW (https://chlorobox.mpimp-golm.mpg.de/cite-OGDraw.html, Lohse et al., 2013) and modified manually. Comparison of plastid genomes of Brunellia antioquensis and B. trianae Comparisons of the boundaries between the large single-copy (LSC) and the small single- copy (SSC) regions with the inverted repeats (IRA and IRB) were performed using Unipro UGENE v1.31.0 (Okonechnikov, Golosova & Fursov, 2012). Content percentages of A, T, C, G, A/T and G/C were estimated for LSC, SSC and IR regions, and for the rRNA, tRNA, protein coding regions, intergenic regions, and introns of both species with Bioedit v7.1.11 (Hall, 1999). Repeated sequences Identification and location of complement, forward, palindromic, and reverse repeat sequences were conducted using REPuter program (Kurtz et al., 2001). Simple sequence repeats (SSRs) were identified using the MISA program with parameters set by default (Thiel et al., 2003). Comparative genome analysis We compared Brunellia plastid genomes regarding nucleotide diversity (π), insertion- s/deletions (InDel) and base substitutions with the program DnaSP6 (Rozas et al., 2017). Transitions (Ts) and transversions (Tv) events were identified across all loci and for protein-coding regions, they were also classified based on the effect in the amino acid chain as synonymous (S) or non-synonymous (N) substitutions. Furthermore,